Signal Loss Decomposition Index (SLDI)

GMI-MKTP-003Marketing Data PrivacyInfrastructure-Dependent

Definition

SLDI provides a three-component decomposition of post-deprecation performance changes into actual performance change, measurement methodology change, and optimization quality change.

The Measurement Gap

After a signal loss event (cookie deprecation, ATT enforcement, a consent rate decline) the marketing dashboard moves. Revenue attribution shifts, channel-level ROAS changes, audience segment performance fluctuates. The immediate question is always: did campaigns actually get worse, or is the measurement system just seeing less? But there is a third possibility that legacy metrics cannot even conceptualize: the AI optimizer degraded because it lost signal quality, and its decisions got worse independently of both campaign performance and measurement accuracy. Legacy metrics like ROAS, CPA, and attributed revenue report a single number that blends all three effects together. They cannot distinguish between "we are performing worse," "we are measuring differently," and "our AI is optimizing poorly because it lost data." Cost per acquisition may rise because actual customer acquisition costs increased, or because the attribution model under-counts conversions it can no longer see, or because the bid optimizer is making worse decisions with degraded signal. Each of these three causes requires a fundamentally different response (budget reallocation, measurement recalibration, or optimizer retraining) but legacy metrics provide no mechanism to determine which cause is dominant. Without decomposing the observed change into its three components, organizations default to treating measurement artifacts as performance problems, wasting budget on campaign optimization when the real issue is signal infrastructure.

Decision Supported:

After a signal loss event moved the dashboard, was the change driven by actual campaign performance, measurement methodology distortion, or degraded AI optimization quality, and therefore which remediation approach should be prioritized?

Formula

Plain language: Break the total observed performance change into three additive pieces: how much came from real performance shifts (measured via incrementality testing), how much came from the measurement system changing (estimated via MMDI), and how much is residual. Attributed to optimization quality degradation. Report each as a percentage of the total observed change.

Notation:

Δ_observed = Δ_performance + Δ_measurement + Δ_optimization

SLDI reports each component as: Component / |Δ_observed| × 100

Variables:

Variable Definition Unit
Δ_observed Total reported change in the target metric across the signal loss event % or currency
Δ_performance The portion of change attributable to actual campaign or market performance, isolated via incrementality testing % or currency
Δ_measurement The portion of change attributable to measurement methodology shifts, estimated via MMDI % or currency
Δ_optimization The residual change after subtracting performance and measurement components, attributed to AI optimization quality degradation % or currency

Proposed formula, pending validation pilots.

At a Glance

Unit
Percentage
Trend Direction
Lower; measurement and optimization components
Reporting Frequency
Quarterly
Category
Signal Degradation and Measurement Integrity
← Modeled Conversion Ratio (MCR)Signal-Quality Allocation Bias (SQAB) →

Assess Your Organization

The AI-GPS Assessment evaluates your AI governance maturity and recommends KPIs like this one for your implementation roadmap.

Take the Free AI-GPS Assessment